INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

PARATHYROID HORMONE-RELATED protein (PTHrP) was first discovered as the causative factor in humoral hypercalcemia of malignancy (33, 47). Although full-length PTHrP can be posttranslationally cleaved into multiple smaller peptides, most of the known biological effects of PTHrP occur via binding of NH₂-terminal-containing peptides to the PTH receptor (PTH1R) (33). Thirty-four amino acid, NH₂-terminal peptides of PTH and PTHrP bind with similar affinity to classic G protein-coupled PTH1R to activate two signaling pathways, adenylyl cyclase/protein kinase A (PKA) and phospholipase C/protein kinase C (PKC) (33, 47). In malignancy, high circulating plasma levels of tumor-derived PTHrP bind to PTH1R in kidney and bone causing hypercalcemia (33, 47). However, subsequent to the discovery of PTHrP and the cloning of PTH1R, it is now appreciated that both PTHrP and PTH1R are expressed in a wide variety of normal tissues and cell types where PTHrP, in contrast to its systemic effects in malignancy, acts locally in a paracrine or autocrine fashion (33, 47).

One such demonstrated site of paracrine or autocrine PTHrP action is the non-central nervous system (CNS) vasculature. In vitro, cytokines or hypoxia induce PTHrP expression in vascular endothelial cells (7, 8, 41), whereas vasoconstrictive peptides induce PTHrP expression in smooth muscle cells (19). Ex vivo PTHrP(1-34) treatment causes vasodilation and increased blood flow in the perfused rat kidney, heart, aorta, or femoral artery and in the human placenta (5, 6, 27, 30, 31). Overexpression of PTH1R in the vascular smooth muscle cells of transgenic mice results in a decrease in systemic blood pressure and peripheral vascular resistance (34), whereas intravenous bolus administration of PTHrP(1-34) can cause transient hypotension (30). To our knowledge, the effects of PTHrP on vascular tone in the CNS have not been explored.

PTHrP and PTH1R are also expressed in the brain. However, the role and regulation of PTHrP in the CNS are not well understood (33). Neurons have been identified as the main site of expression of PTHrP and its cognate receptor in normal brain (33, 48, 49). In vitro studies suggest that neuronal PTHrP may be induced during excitotoxic cell injury or apoptosis and prevent cellular death via autocrine or paracrine binding to the PTH1R (2, 32). Transformed or reactive astrocytes (16, 29, 42), but not normal glial cells, also express PTHrP and PTH1R. In vitro treatment with PTHrP(1-34) induces glial expression of IL-6 (16), a cytokine with neuroprotective effects during cerebral ischemia (26). In contrast, the regulation and function of PTHrP expression in the CNS vasculature have, to our knowledge, not previously been explored.

PTHrP is a member of the cascade of cytokines induced during the inflammatory response in non-CNS organs (10, 15). Recent studies by Funk et al. (16) demonstrating PTHrP induction in reactive astrocytes in response to mechanical brain injury have provided the first evidence that increased PTHrP expression may also accompany CNS inflammation. Of the many factors that can incite an inflammatory response in brain, cerebral ischemia is one of the most clinically important. We therefore postulated that cerebral ischemia might also be a stimulus for increased PTHrP expression in brain. Indeed, hypoxia has already been demonstrated to induce PTHrP expression in non-CNS tissues, such as the kidney and heart (41, 43), and cytokines induced during cerebral ischemia, such as TNF and IL-1, are known to stimulate PTHrP expression in vascular endothelial cells and glia (7, 8, 16). Although the effects of PTHrP(1-34) on cerebral blood flow are not known, we hypothesized that locally produced PTHrP could help maintain blood flow in ischemic brain by causing cerebral vasodilation.

To test these hypotheses, studies were first undertaken to determine whether PTHrP expression was indeed induced in brain in response to ischemic injury using permanent occlusion of the middle cerebral artery (MCA) in rats as an experimental model. The time course of increased PTHrP expression was compared with that of other inflammatory cytokines known to be activated by ischemic stroke (1). To determine whether the cerebral microcirculation is also a potential target for PTHrP action in brain, vasodilatory effects of PTHrP on pial arterioles, vessels that mirror the vascular response of deeper cortical vessels (39), were assessed by intravital fluorescence microscopy. Finally, to elucidate a possible protective function of PTHrP during cerebral ischemia, the effect of intracerebroventricular PTHrP(1-34) administration on infarct size was determined in MCA-occluded (vs. sham) animals.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Materials PTHrP(1-34) and PTH(3-34) were obtained from Bachem (Torrance, CA); human serum albumin (HSA) was obtained from Immuno-U.S. (Rochester, MI); and forskolin and barium chloride were obtained from Sigma (St. Louis, MO).

MCA occlusion. All experimental animal procedures were conducted in accordance with University of Arizona and Institute for Laboratory Animal Research guidelines using anesthetized male Sprague-Dawley rats (250-350 g). MCA occlusion (MCAO) was induced using the intraluminal filament method, as previously described by Ritter et al. (36, 40). Briefly, rats were anesthetized via a facemask with 1 liter of N₂O, 0.5 liter of O₂, and 0.5 to 1% halothane while maintaining a constant body temperature (37 ± 0.5°C). After exposure of the right common carotid and cauterization of appropriate branches, the external carotid artery was isolated and cauterized. A nylon filament (3-0) with a rounded tip was inserted into the external carotid stub and advanced 18 mm into the internal carotid artery. The filament was secured, the neck incision was sutured, and the animals were allowed to recover. Sham-operated animals underwent the same surgical procedure, excluding placement of the intraluminal filament. One hour after permanent filament placement or sham operation, neurological function was assessed using four standard tests, as previously described by Ruehl et al. (40): 1) level of consciousness (LOC), 2) spontaneous circling, 3) front limb paresis, and 4) front limb symmetry. To be included in the study, MCAO animals had to demonstrate a minimum score of 1 in every test (scored 0-4/test), a total minimum score of 6, and a maximum score of 3 in the LOC test (absence of coma or seizures).

Northern analysis. Brains were removed from MCAO and sham-operated animals 2, 4, 12, or 24 h following ischemic injury and quickly dissected coronally to discard the noninjured frontal cortex and cerebellum. The remaining hemispheres were then bisected longitudinally and frozen separately in liquid nitrogen before storage at 70°C. Polyadenylated RNA, isolated from the hemispheres ipsilateral or contralateral to MCAO (or sham operation), was assessed by Northern analysis using methods and ³²P-labeled cDNA probes previously described by Funk et al. (11-13, 16) to determine the time course of changes in expression of mRNA for PTHrP and its cognate receptor (PTH/PTHrP receptor) vs. other CNS inflammatory cytokines known to be activated by ischemia (TNF-, IL-1, and IL-6) (1). Blots were exposed to film at 70°C using intensifying screens, and autoradiographic intensity was quantitated using a BioRad model GS-700 Imaging Densitometer.

Immunohistochemistry. Brains were immediately removed from euthanized animals, cut into 2-mm coronal sections, and fixed in 10% buffered formalin before paraffin embedding. Serial tissue sections were processed for immunohistochemical staining as previously described by Funk et al. (11) using an affinity-purified polyclonal primary antibody directed against PTHrP(34-53) (Oncogene, Cambridge, MA). Brain sections were processed for antigen unmasking by heating for 10 min at 95°C in 10 mmol/l sodium citrate (pH 6.0) before immunostaining. Specificity of PTHrP immunostaining was verified by the absence of staining observed when serial tissue sections were treated with PTHrP antibody that had been preincubated with a 20-fold excess by weight of PTHrP(34-53) peptide (Oncogene). Astrocytes, endothelial cells, and smooth muscle cells were also identified in serial sections using antibodies directed against glial fibrillary acidic protein (GFAP; Zymed, South San Francisco, CA), Factor VIII-related antigen (BioGenex, San Ramon, CA), or smooth muscle actin (Zymed), respectively.

Cranial window placement and intravital microscopy preparation. Male Sprague-Dawley rats (250 to 350 g) were anesthetized with pentobarbitol sodium (50 mg/kg), intubated, and continuously respirated. Anesthesia was maintained with 0.1 ml (50 mg/ml) of pentobarbital sodium per hour as needed. Tail artery and vein catheters were placed for continuous measurement of blood pressure and administration of drugs, respectively. To achieve muscle paralysis for a stable microscopic field, vecuronium bromide (2.4 mg/h) was infused continuously. Throughout each experiment, body temperature was monitored continuously with a rectal probe and maintained at 37°C with a heating pad. An open cranial window was prepared over the right cortical-parietal brain surface as previously described by Ritter et al. (36). Immediately after the craniotomy, mineral oil was placed over the preparation to prevent exposure of the pial vasculature to air during removal of the dura. Subsequent to removal of the dura, the microvascular preparation was continuously superfused with a 37°C artificial cerebrospinal fluid (aCSF) solution that was monitored with respect to gas tensions and pH (Radiometer, ABL) (40). With the use of fluorescence videomicroscopy and intravenous injection of FITC-BSA (1 ml of a 5% solution) to identify arteriole margins, changes in cerebral arteriole diameter were assessed in randomly selected pial arterioles by measurement of TV monitor images using a calibrated ruler. In each experiment, peptide superfusion was preceded by superfusion with peptide vehicle alone (0.1% apyrogenic HSA in aCSF) to monitor for nonspecific effects of vehicle on arteriole tone. Additionally, the vascular reactivity of all pial preparations was verified 15 min after the peptide superfusions by treatment with 10⁶ mol/l forskolin, a cAMP-stimulating agent that is a known cerebral arteriolar vasodilator (45), followed by 2.5% barium chloride (BaCl₂), a known vasoconstrictor (38). Each pial arteriole preparation was superfused once with PTHrP(1-34), delivered at a constant rate by mixing with continuously flowing aCSF (1.7 ml/min) to achieve a final PTHrP concentration of 1 × 10⁶ mol/l. This dose elicits a maximal dilatory and/or blood flow response in non-CNS vascular beds (5, 6, 27). In one set of experiments, animals were treated with 1 × 10⁶ mol/l PTH(3-34), a peptide that does not activate the adenylyl cyclase pathway (31) before superfusion with PTHrP(1-34) to identify postreceptor pathways responsible for the vascular effects of PTHrP.

Intracerebral ventricular infusion of PTHrP(1-34). One week before MCAO, outer guide cannulas (Plastics One, Roanoke, VA) were placed in pentobarbital sodium-anesthetized rats under stereotaxic control using coordinates and methods previously described by Chen et al. (4, 35) to allow for later placement of an inner cannula into the right lateral ventricle. At the time of the experiment, animals were prepared as described for MCAO, with the additional manipulation of placement of the inner cannula into the lateral ventricle (4, 35, 37). For intracerebroventricular PTHrP(1-34) peptide treatment, a dosing strategy similar to that successfully used for intracerebroventricular treatment of stroke with other short-acting peptides, such as IL-6 and IL-1ra, was employed (25, 26). Animals were given a bolus (5 µl/1-2 min) of 200 ng PTHrP(1-34) in apyrogenic 0.1% HSA in normal saline (NS) vs. 0.1% HSA/NS alone (vehicle) 30 min before and 90 min after permanent occlusion of the MCA. Because up to 50% of the total amount of intracerebroventricularly administered peptides can appear in the peripheral circulation over 4 h (3), a dose of PTHrP (200 ng) was chosen that could not elicit a systemic hypotensive response (30). Arterial blood pressure was continuously monitored by arterial line at the times of injections. Subsequent to the last intracerebroventricular injection, tail catheters and inner intracerebroventricular cannulas were removed. Animals were allowed to recover with neurological testing at 4 h and removal of brain for infarct analysis at 24 h.

Measurement of cerebral infarction size. Cerebral infarction volume was determined as previously described by Ruehl et al. (40) using standard methods. Briefly, 24 h after MCAO, animals were euthanized by an overdose of halothane. The brains were immediately removed and sectioned into seven 2-mm coronal slices, followed by immersion in 2% triphenyl tetrazolium chloride and subsequent fixation in 10% buffered formalin. For image analysis, each brain section was photographed, scanned (600 dpi), and the area of infarction and area of each hemisphere were measured using National Institutes of Health Image software. The contribution of edema (which was not different between controls and treated animals) to infarct volume was corrected for using standard methods (by subtracting the volume of the noninfarcted ipsilateral hemisphere from the volume of the contralateral hemisphere), with infarcted volume expressed as a percent of the contralateral hemisphere (40).

Statistical analysis. Values are presented as means ± SE with statistical significance determined by ANOVA with post hoc testing, Student's t-test, or Mann-Whitney testing (Instat, Graphpad, San Diego, CA). For analysis of changes in vessel diameter over time, values for an individual vessel were compared by paired analysis.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effect of MCAO on cerebral PTHrP and cytokine mRNA levels. Twenty-four hours after permanent MCAO, PTHrP mRNA levels were increased fourfold in the ischemic cortex compared with the nonischemic contralateral cortex (Fig. 1, A and B). In sham-operated animals, PTHrP mRNA levels were unchanged and no different than those found in the nonischemic cortex of MCAO animals (Fig. 1B). At 24 h, mRNA levels for TNF-, IL-1, and IL-6 were also increased four- to eightfold in the ischemic (vs. contralateral nonischemic) cortex (Fig. 1C). Unlike PTHrP, mRNA levels for these cytokines were not detected by Northern analysis in sham-operated animals (Fig. 1C). However, consistent with previous reports (23, 24), TNF-, IL-1, and IL-6 mRNA levels were detectable, albeit at lower levels, in the hemisphere contralateral to the ischemic injury (Fig. 1C). The time course of mRNA induction for these cytokines was compared with PTHrP in the ischemic cortex (Fig. 2). As has been previously reported (23, 24), TNF- and IL-1 mRNA levels were increased at the earliest time point examined (2 h after permanent MCAO) (Fig. 2, A and B). In contrast, PTHrP and IL-6 mRNA levels in the ischemic cortex were not increased until 4 h after MCAO (Fig. 2, C and D). PTH/PTHrP receptor mRNA levels, which were unchanged in sham-operated animals, exhibited a small, but statistically significant, decrease in ischemic (vs. contralateral nonischemic) brain at 24 h (Fig. 1B), but not at earlier times (data not shown).

View larger version (32K): [in this window] [in a new window]   Fig. 1.   Effect of 24-h permanent middle cerebral artery occlusion (MCAO) (vs. sham operation) on parathyroid hormone-related protein (PTHrP), PTH receptor (PTH1R), and cytokine mRNA expression. A: Northern analysis of PTHrP (and actin) mRNA levels in ischemic vs. nonischemic hemispheres 24 h after MCAO, performed as described in MATERIALS AND METHODS. B: PTHrP and PTH/PTHrP receptor mRNA levels in cerebral hemispheres ipsilateral (+) or contralateral () to MCAO (or sham operation) are expressed in arbitrary densitometry units relative to actin (n = 1 sham, or means ± SE for n = 3 MCAO animals). C: TNF-, IL-1, and IL-6 mRNA levels were similarly determined in the same animals as in B. B and C: differences between ischemic and nonischemic contralateral cortex in MCAO animals are statistically significant by Student's t-test. *P < 0.05 vs. contralateral hemisphere. ND, not detected.

View larger version (54K): [in this window] [in a new window]   Fig. 2.   Time course of induction of PTHrP and cytokine mRNA expression following permanent MCAO. TNF- (A), IL-1 (B), PTHrP (C), and IL-6 (D) mRNA levels in ischemic vs. nonischemic cerebral hemispheres were determined at the indicated times after permanent MCAO. Results are expressed in arbitrary densitometry units relative to actin (means ± SE for n = 3 animals). Where indicated, differences between ischemic and nonischemic contralateral cortex in MCAO animals are statistically significant by Student's t-test. *P < 0.05 vs. contralateral hemisphere. **P < 0.01 vs. contralateral hemisphere.

Immunohistochemical localization of PTHrP following focal stroke. PTHrP protein was localized in normal and ischemic brain by immunohistochemical analysis (Fig. 3). Specificity of PTHrP staining was verified in all cases by the absence of staining seen in consecutive tissue sections treated with PTHrP antibody that had been preincubated with an excess of antigen (e.g., Fig. 3E vs. 3F). In nonischemic brain, immunoreactive PTHrP in the striatum (Fig. 3A) and cortex (Fig. 3B) was located in neurons (double arrowheads), while the microvasculature (arrows) was PTHrP negative. Four hours following permanent MCAO, neuronal PTHrP immunoreactivity persisted and PTHrP immunoreactivity also began to appear in the vessels of the ischemic hemisphere (data not shown). At 24 h, neuronal PTHrP staining (double arrowheads) persisted in the infarct penumbra (Fig. 3C) but decreased in the infarct [Fig. 3D (cortex)/3E (striatum)], whereas vascular PTHrP staining increased in large and small vessels in all areas of injury (Fig. 3, C-E, arrows). Astrocytes showed no PTHrP immunoreactivity 4 or 24 h after MCAO. Comparison of immunostaining for vascular PTHrP (Fig. 3, C-E), Factor VIII-related antigen-positive endothelial cells (Fig. 3G), smooth muscle actin-positive vascular smooth muscle cells (Fig. 3H), and GFAP-positive perivascular astrocytes (Fig. 3I) in serial sections suggested that longitudinal endothelial cells lining the vasculature, and not circumferential vascular smooth muscle cells or perivascular astrocytes, were the source of immunoreactive PTHrP in ischemic vessels.

View larger version (132K): [in this window] [in a new window]   Fig. 3.   Immunohistochemical localization of PTHrP following permanent MCAO. A: in the striatum of sham-operated animals, immunoreactive PTHrP (brown) was primarily located in neurons (double arrowheads) in a fine reticular pattern in the perikarya, whereas the vasculature (arrows) was PTHrP negative. B: cortical PTHrP immunoreactivity was similarly located, in a more diffuse pattern, in neuronal perikarya of sham-operated animals (double arrowheads), whereas the vasculature (arrows) was PTHrP negative. C: after MCAO (24 h), PTHrP persisted in the neurons in the infarct penumbra (double arrowhead), whereas vascular endothelial cells (arrows) became PTHrP positive. D: in the infarcted cortex, the vasculature (arrows) was similarly PTHrP positive, whereas neuronal staining (double arrowheads) persisted but was somewhat decreased. E: within the infarcted striatum, vascular endothelial cells also became PTHrP positive (arrows), whereas immunoreactivity of striatal neurons was significantly decreased. F: specificity of PTHrP staining was verified by the absence of staining seen when PTHrP(34-53) antibody was preincubated with antigen, as indicated here in a serial section of the striatum shown in E. G: Factor VIII-related antigen immunostaining of endothelial cells lining ischemic vessels (arrows) revealed a staining pattern similar to that of PTHrP. H: in contrast, smooth muscle actin-positive vascular smooth muscle cells (arrows) formed a concentric ring around the longitudinal endothelial cells (double arrowhead) in larger vessels and were absent from small PTHrP-positive capillaries (not shown). I: foot processes of glial fibrillary acidic protein-positive astrocytes (arrows) circumscribe the perimeter of the vascular cells. Nuclei are counterstained with methyl green.

Effects of PTHrP(1-34) on pial arteriolar diameter. When the pial microcirculation of normal animals was sequentially superfused (5 min/drug with a 5- to 15-min aCSF washout between drugs) with vehicle (0.1% HSA), PTHrP(1-34) (10⁶ mol/l), forskolin (10⁶ mol/l), and BaCl₂ (2.5%), PTHrP(1-34) and forskolin both caused a significant increase in arteriolar diameter (31 and 54%, respectively), whereas BaCl₂ caused arteriolar diameter to decrease (Fig. 4A). Comparison of the dilatory response of larger (>25 µm) vs. smaller (<25 µm), more terminal arterioles revealed a dilatory effect of PTHrP and forskolin, a cAMP-stimulating agent, on vessels of both sizes (Fig. 4B). However, the smaller terminal arterioles had a greater dilatory response to both PTHrP(1-34) (42%) and forskolin (76%) (Fig. 4B). Prolonged treatment with PTHrP(1-34) resulted in an immediate (1 min) and persistent increase in arteriolar diameter over 25 min of superfusion (Fig. 4C). Vessels remained responsive to subsequent challenge with 10⁶ mol/l forskolin (43% increase, P < 0.001) following prolonged PTHrP(1-34) treatment. Superfusion with 10⁶ mol/l PTH(3-34), a peptide that binds the PTH/PTHrP receptor but does not activate the cAMP signaling pathway (31), had no effect on arteriolar diameter (Fig. 4D, ), while subsequent challenge with 10⁶ mol/l PTHrP(1-34) elicited a typical 30% increase in arteriolar diameter (Fig. 4D, ). A representative example of the vasodilatory response of the pial microcirculation to PTHrP(1-34) is given in Fig. 4E (baseline) and Fig. 4F (during PTHrP[1-34]). Mean arterial blood pressure did not change in response to superfusion with PTHrP(1-34) peptide, even after 25 min of treatment (data not shown). Physiological PaO₂ (132.2 ± 7.7 mmHg) and PaCO₂ (32.1 ± 1.10 mmHg) values were carefully maintained throughout the course of the experiments.

View larger version (56K): [in this window] [in a new window]   Fig. 4.   Effect of PTHrP(1-34) on pial arteriole diameter. A: pial microcirculation was sequentially superfused (5 min/solution), as described in MATERIALS AND METHODS, with vehicle [0.1% human serum albumin (HSA) in artificial cerebrospinal fluid (aCSF)], 1 × 106 mol/l PTHrP(1-34), 1 × 106 mol/l forskolin, or 2.5% barium chloride, with 5- to 15-min aCSF washouts between each treatment. Results are expressed as the maximal percent change in arteriole diameter (means ± SE for n = 23 vessels observed in 6 animals) during the 5 min of superfusion for each solution. Statistically significant changes in arteriole diameter from baseline (baseline diameter, 26.9 ± 1.0 µm) were determined by paired Student's t-test for each treatment. ***P < 0.001 vs. baseline. B: effects of 5 min of sequential superfusion with 0.1% HSA, 1 × 106 mol/l PTHrP(1-34), or 1 × 106 mol/l forskolin on arteriole diameter, expressed as fold change relative to baseline, are presented for vessels (n = 11-12/group) <25 µm (average diameter 19.7 ± 1.2 µm) or >25 µm (average diameter 34.8 ± 1.8 µm). Statistically significant changes in arteriole diameter from baseline were determined by paired Student's t-test for each treatment. ***P < 0.001 vs. baseline. Differences between treatments on fold change in arteriole diameter were determined by ANOVA. **P < 0.001. C: pial microcirculation was continuously superfused for 25 min with 1 × 106 mol/l PTHrP(1-34) and effects on arteriole diameter, expressed as fold change relative to baseline, were determined every 1-2 min (n = 20 vessels in 4 animals). Changes in arteriole diameter relative to baseline (baseline diameter 27.3 ± 2.2 µm) were determined by paired ANOVA. *P < 0.05. **P < 0.01. ***P < 0.001. D: pial microcirculation was sequentially superfused (5 min/solution) with 1 × 106 mol/l PTH(3-34), followed by 1 × 106 mol/l PTHrP(1-34), with an intervening aCSF washout. Effects of treatment on changes in arteriole diameter (means ± SE, n = 25 vessels in 4 animals) relative to baseline were determined by paired ANOVA. ***P < 0.001. E: representative pial arterioles at baseline before PTHrP(1-34) infusion. F: same pial arterioles as in E 17 min after start of PTHrP(1-34) infusion.

Effects of PTHrP(1-34) treatment on infarct size. Physiological parameters, including mean arterial blood pressure, pH, PaCO₂, temperature, and glucose, were the same in vehicle- and PTHrP(1-34)-treated animals at baseline, immediately pre-MCAO, and pre- and posttreatments. No experimental animals were excluded from either treatment group based on neurologic scoring. Mean total infarct size in rats treated with PTHrP(1-34) peptide was 34% less than that in animals treated with vehicle alone (Fig. 5), but this difference was not significant (P < 0.06). However, examination of infarct volume in the cortical vs. subcortical areas (Fig. 5) demonstrated a significant reduction in cortical infarct volume in the PTHrP(1-34)-treated group (47%, P < 0.02). Subcortical infarct size was not affected by PTHrP(1-34) peptide treatment (Fig. 5).

View larger version (30K): [in this window] [in a new window]   Fig. 5.   Effect of intracerebroventricular PTHrP(1-34) peptide on infarct volume following permanent MCAO. Animals were treated intracerebroventricularly with PTHrP(1-34) peptide (vs. vehicle alone) 30 min before and 90 min post MCAO. Brains were harvested 24 h after permanent MCAO for analysis of infarct volume, as described in MATERIALS AND METHODS. Total infarct volume, cortical infarct volume, or subcortical infarct volume is expressed as percent of the contralateral hemisphere (means ± SE, n = 16/group). Effect of PTHrP(1-34) on infarct volume was determined by 2-tailed Mann-Whitney testing. *P < 0.02. ns, Not significant.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

These experiments provide, to our knowledge, the first evidence that enhanced PTHrP gene expression is induced in the brain in response to focal ischemia. Unlike other inflammatory cytokines, PTHrP is expressed constitutively in brain, mainly by neurons, including those in the parietal cortex and striatum (48, 49). Because increased CNS PTHrP expression was previously described in reactive astrocytes formed in response to stab wound injury (16), we anticipated finding increased PTHrP expression in reactive astrocytes in the infarct penumbra in response to CNS ischemia. Instead, the vasculature of the injured hemisphere, rather than reactive astrocytes, was found to be the site of increased immunoreactive PTHrP during the first 24 h following permanent MCAO.

Just as endothelial cells have been reported to be the source of increased PTHrP expression in the ischemic myocardium (41), vascular endothelial cells appeared to be the source of increased immunoreactive PTHrP in the vasculature of ischemic brain. Increased immunoreactive PTHrP was found in vascular endothelial cells as early as 4 h after MCAO and persisted for up to 24 h. This increase in vascular PTHrP protein may be the result of a local increase in gene expression, as PTHrP mRNA levels were increased in the ischemic hemisphere over the same time period. Additionally, preliminary evidence of a positive arteriovenous PTHrP gradient across the ischemic brain at 24 h (i.e., 20% lower PTHrP levels in superior sagittal sinus vs. aortic plasma; Funk and Ritter, unpublished data) before breakdown of the blood-brain barrier (17, 22) suggests the possibility that uptake of PTHrP from the circulation may also contribute to the increase in vascular PTHrP demonstrated at later time points in ischemic brain (35, 44).

PTHrP mRNA induction in the ischemic hemisphere was preceded by induction of mRNA for TNF- and IL-1, two cytokines that have been demonstrated to induce PTHrP expression in other in vivo and in vitro models of inflammation, including endotoxemia and cytokine stimulation of endothelial cells and astrocytes (7, 8, 12, 16, 42). This finding is therefore consistent with the postulate that TNF- and/or IL-1 may also mediate ischemia-induced PTHrP expression in the brain. Similarly, because PTHrP can induce IL-6 expression in multiple cell types, including glia (11, 16), the delayed expression of IL-6 found in ischemic brain might also be attributable to local increases in PTHrP.

At all time points examined, mRNA for the PTH/PTHrP receptor (PTH1R) was expressed in the ischemic hemisphere, although at the time of maximal induction of PTHrP (24 h), levels of receptor expression were decreased. This reciprocal regulation of PTHrP and PTH1R, which is consistent with the well-described ability of PTHrP to downregulate the expression of its receptor (9, 14, 43), provides further evidence of a biological effect of locally enhanced PTHrP expression in ischemic brain.

Given our finding of increased immunoreactive PTHrP in the microcirculation of the ischemic brain, we postulated that one possible protective effect of this vasoactive peptide during stroke could be to enhance cerebral blood flow. Consistent with this hypothesis and with the known vasodilatory effects of NH₂-terminal PTHrP in non-CNS vascular beds (5, 6, 27, 30, 33), we found that superfusion of the pial microcirculation with PTHrP(1-34) significantly increased arteriole diameter by 30%. According to Poiseuille's equation, wherein flow is proportional to the fourth power of the vessel radius, this 30% increase in arteriolar diameter could result in a threefold increase in blood flow in a setting of constant pressure, as was documented in these experiments. Because the smaller terminal arterioles, such as those studied here, strongly influence cerebrovascular resistance and blood flow (18), PTHrP may therefore play a critical role in the maintenance of cerebrovascular blood flow. Moreover, because the vascular responses of the pial arterioles are similar to the cerebral circulation as a whole and because changes in pial arteriole diameter parallel changes in regional blood flow (39), these findings suggest that ischemia-induced PTHrP in microvessels in areas of ischemic brain could serve to enhance cerebral blood flow to the damaged cortex.

Binding of NH₂-terminal PTH/PTHrP peptides to the PTH1R can stimulate adenylyl cyclase/PKA and/or phospholipase C/PKC signaling pathways (33, 47). A previous report by Huang et al. (21) demonstrated a PTH(1-34)-mediated increase in cAMP formation in cerebral microvessels ex vivo, suggesting that the vasodilatory effects of PTHrP(1-34) demonstrated here could be mediated via an adenylyl cyclase signaling pathway. Consistent with this hypothesis and with the demonstrated role of cAMP in mediating PTHrP(1-34) vasodilation in non-CNS vascular beds (28, 46), superfusion of the pial arterioles with PTH(3-34), a peptide that binds to the PTH1R but does not stimulate cAMP formation (21, 31), had no effect on arteriolar diameter. Homologous desensitization to the sustained vasodilatory effects of PTH/PTHrP peptides, but not heterologous desensitization to subsequent dilation by other cAMP-stimulating agents, such as forskolin, has been reported to occur in response to PTH/PTHrP peptides in some non-CNS vascular beds (28, 31). However, under the conditions of the experiments described here, neither homologous nor heterologous desensitization of the cerebral microcirculation was seen in response to PTHrP(1-34) treatment. Because parenchymal and pial arterioles respond similarly to vasoactive stimuli (39), these findings suggest that sustained increases in PTHrP in the cerebral microcirculation, such as those occurring during ischemia, may be associated with a sustained increase in the diameter of those microvessels that regulate local blood flow.

Finally, the demonstration of a protective effect of PTHrP(1-34) peptide treatment in limiting cortical infarct size is consistent with the hypothesis that endogenously produced PTHrP also has a protective effect in ischemic brain. In particular, because a lack of collateral blood flow and differences in microvascular structure that may allow for early plugging of terminal arterioles make the striatal region more difficult to salvage following MCAO (50), the isolated protective effect of PTHrP(1-34) treatment in decreasing cortical, but not striatal, infarct size suggests that this vasodilatory peptide is neuroprotective in those areas of the brain that can be most easily salvaged by an increase in blood flow.

In summary, the studies described here provide novel evidence demonstrating an increase in local vascular PTHrP gene expression in ischemic brain, as well as the ability of NH₂-terminal PTHrP to act as a potent vasodilator in the pial microcirculation and to reduce cortical infarct size by almost 50%. Additional studies will be required to identify all possible CNS targets for PTHrP action during ischemia, as PTHrP, in addition to the vasodilatory effects demonstrated here, has also been reported to have direct protective effects on neurons and to induce glial expression of neuroprotective cytokines (2, 16, 32). However, the beneficial effect of PTHrP on cortical infarction is consistent with the hypothesis that endogenous increases in cerebral PTHrP may serve to protect the brain during ischemia by preserving cerebral blood flow. Moreover, the demonstration of a protective effect of exogenously administered PTHrP(1-34) suggests that this peptide, which has been administered in clinical trials for the treatment of osteoporosis in doses as high as 400 µg/day (20), may also be useful in acute therapeutic interventions aimed at improving clinical outcomes in patients suffering from stroke.